Applications of Deep Learning in Molecule Generation and Molecular Property Prediction

被引:151
|
作者
Walters, W. Patrick [1 ]
Barzilay, Regina [2 ]
机构
[1] Relay Therapeut, Cambridge, MA 02142 USA
[2] MIT, Comp Sci & Artificial Intelligence Lab, 77 Massachusetts Ave, Cambridge, MA 02139 USA
关键词
TRAINING SET; PUBCHEM; DESIGN; DATABASE; DOMAIN;
D O I
10.1021/acs.accounts.0c00699
中图分类号
O6 [化学];
学科分类号
0703 ;
摘要
CONSPECTUS: Recent advances in computer hardware and software have led to a revolution in deep neural networks that has impacted fields ranging from language translation to computer vision. Deep learning has also impacted a number of areas in drug discovery, including the analysis of cellular images and the design of novel routes for the synthesis of organic molecules. While work in these areas has been impactful, a complete review of the applications of deep learning in drug discovery would be beyond the scope of a single Account. In this Account, we will focus on two key areas where deep learning has impacted molecular design: the prediction of molecular properties and the de novo generation of suggestions for new molecules. One of the most significant advances in the development of quantitative structure-activity relationships (QSARs) has come from the application of deep learning methods to the prediction of the biological activity and physical properties of molecules in drug discovery programs. Rather than employing the expert-derived chemical features typically used to build predictive models, researchers are now using deep learning to develop novel molecular representations. These representations, coupled with the ability of deep neural networks to uncover complex, nonlinear relationships, have led to state-of-the-art performance. While deep learning has changed the way that many researchers approach QSARs, it is not a panacea. As with any other machine learning task, the design of predictive models is dependent on the quality, quantity, and relevance of available data. Seemingly fundamental issues, such as optimal methods for creating a training set, are still open questions for the field. Another critical area that is still the subject of multiple research efforts is the development of methods for assessing the confidence in a model. Deep learning has also contributed to a renaissance in the application of de novo molecule generation. Rather than relying on manually defined heuristics, deep learning methods learn to generate new molecules based on sets of existing molecules. Techniques that were originally developed for areas such as image generation and language translation have been adapted to the generation of molecules. These deep learning methods have been coupled with the predictive models described above and are being used to generate new molecules with specific predicted biological activity profiles. While these generative algorithms appear promising, there have been only a few reports on the synthesis and testing of molecules based on designs proposed by generative models. The evaluation of the diversity, quality, and ultimate value of molecules produced by generative models is still an open question. While the field has produced a number of benchmarks, it has yet to agree on how one should ultimately assess molecules "invented" by an algorithm.
引用
收藏
页码:263 / 270
页数:8
相关论文
共 50 条
  • [1] Advanced deep learning methods for molecular property prediction
    Pang, Chao
    Tong, Henry H. Y.
    Wei, Leyi
    [J]. QUANTITATIVE BIOLOGY, 2023, 11 (04) : 395 - 404
  • [2] Pretraining deep learning molecular representations for property prediction
    Liu, Bowen
    Hu, Weihua
    Leskovec, Jure
    Liang, Percy
    Pande, Vijay
    [J]. ABSTRACTS OF PAPERS OF THE AMERICAN CHEMICAL SOCIETY, 2019, 258
  • [3] Deep learning methods for molecular representation and property prediction
    Li, Zhen
    Jiang, Mingjian
    Wang, Shuang
    Zhang, Shugang
    [J]. DRUG DISCOVERY TODAY, 2022, 27 (12)
  • [4] Deep contrastive learning of molecular conformation for efficient property prediction
    Yang Jeong Park
    HyunGi Kim
    Jeonghee Jo
    Sungroh Yoon
    [J]. Nature Computational Science, 2023, 3 : 1015 - 1022
  • [5] Evidential Deep Learning for Guided Molecular Property Prediction and Discovery
    Soleimany, Ava P.
    Amini, Alexander
    Goldman, Samuel
    Rus, Daniela
    Bhatia, Sangeeta N.
    Coley, Connor W.
    [J]. ACS CENTRAL SCIENCE, 2021, 7 (08) : 1356 - 1367
  • [6] Deep contrastive learning of molecular conformation for efficient property prediction
    Park, Yang Jeong
    Kim, Hyungi
    Jo, Jeonghee
    Yoon, Sungroh
    [J]. NATURE COMPUTATIONAL SCIENCE, 2023, 3 (12): : 1015 - +
  • [7] Explainable uncertainty quantifications for deep learning-based molecular property prediction
    Yang, Chu-, I
    Li, Yi-Pei
    [J]. JOURNAL OF CHEMINFORMATICS, 2023, 15 (01)
  • [8] GMPP-NN: a deep learning architecture for graph molecular property prediction
    Abbassi, Outhman
    Ziti, Soumia
    Belhiah, Meryam
    Lagmiri, Souad Najoua
    Seghroucheni, Yassine Zaoui
    [J]. DISCOVER APPLIED SCIENCES, 2024, 6 (07)
  • [9] Explainable uncertainty quantifications for deep learning-based molecular property prediction
    Chu-I Yang
    Yi-Pei Li
    [J]. Journal of Cheminformatics, 15
  • [10] A general optimization protocol for molecular property prediction using a deep learning network
    Chen, Jen-Hao
    Tseng, Yufeng Jane
    [J]. BRIEFINGS IN BIOINFORMATICS, 2022, 23 (01)